[PATCH] D74661: AMDGPU/GlobalISel: Insert readfirstlane on SGPR returns

[Nicolai Hähnle via Phabricator via llvm-commits]

nhaehnle added a comment.

Come to think of it, is there an explicit test that we *don't* insert v_readfirstlane when none is needed at all?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D74661/new/

https://reviews.llvm.org/D74661